1. Field of the Invention
The present invention relates to a redundant failover system, a redundancy managing apparatus and an application processing apparatus, and particularly to a redundant failover system for performing redundant failover of an application processing apparatus in a communication system in which a plurality of application processing apparatuses cooperate with each other to perform communication control, a redundancy managing apparatus and an application processing apparatus.
2. Description of the Related Art
In an independent wireless system for disaster prevention or fire-fighting use, failure resistance against disaster or the like is very important. Thus, an inter-site redundancy technique as disclosed in, for example, non-patent document 1 is put to practical use. In the technique of non-patent document 1, circuit switches of a site A and a site B, which are geographically separated from each other, have one-to-one redundant structure. It is assumed that the circuit switch at the site A is in an active state, and the circuit switch at the site B is in a standby state. Each site contains lines from peripheral base stations, and in the site B, the line from the base station is connected to a relay line to the site A, not the circuit switch of the site B. The circuit switch of the site A controls all base stations in the independent wireless system. Here, it is assumed that an earthquake occurs at the site A, and a failure occurs in the circuit switch of the site A. In the site B, the line from the base station is connected to the circuit switch of the site B, not the relay line. In the site A, the line from the base station is connected to the relay line to the site B, not the circuit switch of the site A. The circuit switch of the site B is placed in the active state, and all the base stations in the independent wireless system are controlled in the site B.
On the other hand, as a failure resistance technique in an IP-based communication system, for example, there is a Linux Virtual Server (LVS). In the LVS, requests from an unspecified number of Web clients are processed by Web servers dispersed and arranged in a plurality of subnets. First, a Web client inquires of a DNS server about an IP address of URL which the client desires to access. The DNS server sends back the IP address of a load balancer. The Web client transmits an HTTP request packet to the IP address of the load balancer. The load balancer selects the Web server to process the HTTP request by the Hash value of the source IP address of the HTTP request packet. Then, the received packet is encapsulated by a packet addressed to the selected Web server and is transmitted. In all the Web servers, the IP address of the load balancer is set (that is, made effective) as one of IP addresses of the Web server itself. The Web server having received the encapsulated packet decapsulates the packet. Then, since the packet addressed to the IP address of the load balancer, that is, the packet addressed to the Web server appears, an application on the Web server processes the HTTP request. The application on the Web server transmits an HTTP response packet to the IP address of the Web client set in the source address of the packet. The source IP address is the IP address of the load balancer.
As described above, in the LVS, since an IP tunnel is used for the packet transfer from the load balancer to the Web server, the load balancer and the Web server, and the Web servers may be located in different subnets. Besides, the load server periodically transmits an HTTP request to each Web server in order to detect the failure of the Web server. When a response is not returned, the HTTP request from the Web client is made not to be transferred to the Web server. By doing so, also with respect to the Web client whose HTTP request has been transferred to the failed Web server, the HTTP request is transferred to a normal Web server.
Besides, as a technique in which an IP tunnel is used for communication between apparatuses, and an application uses a virtual IP address different from an IP address of the apparatus, there is a technique disclosed in, for example, non-patent document 2. According to the technique of non-patent document 2, a plurality of users on an IP network form a group and secure communication is performed in the group. Member information of the group and a virtual IP address space are previously registered in a DNS server. The user starts the login client in a user terminal, and transmits a login request to the DNS server. The DNS server identifies the group to which the user terminal belongs, and assigns a virtual IP address used by an application of the user terminal. When the application of the user terminal communicates with an application of a terminal of the same group, a packet is transmitted to the virtual IP address of the communication partner application. When it is not known that the destination address of the packet transmitted by the application is a virtual IP address on which terminal, the IPsec driver of the user terminal inquires of the DNS server, and knows the IP address of the communication partner terminal. An IPsec tunnel is established between the user terminal and the communication partner terminal, and the packet transmitted by the application is sent to the communication partner terminal.
[Non-patent document 1] “Large-Scale Fire-Fighting/Emergency Digital Wireless System” Hitachi Kokusai Denki Giho, 2006, No. 7
[Non-patent document 2] “Overlay Network Control System Suitable For Large-Scale Dynamic Group Communication”, Denshi Jouhou Tsusin Gakkai, Singaku Giho, NS2004-95, 2004-09